From 2 Leads a Week to 15: A Roofer's AI Outreach Story

Two Leads a Week Isn't a Pipeline. It's a Gamble.

Kyle Harmon ran a four-person roofing crew in Murfreesboro, Tennessee. Solid work, good reviews, real local reputation. The problem wasn't quality — it was volume.

Two leads a week from Angi. That's what three years and roughly $50,000 in total platform spend had built him. Two shots per week at a job, shared with two or three other roofers who got the same lead notification at the same time.

His crew could handle 12-15 jobs a month. They were running at half capacity. Kyle kept waiting for business to pick up on its own. It didn't.

What Angi Was Actually Costing Him

The monthly fee wasn't the real problem. The real problem was what that money bought.

Here's what Kyle's account looked like in Q4 2024:

• Monthly fee: $1,400
• Leads received: 16-18 per month ($78-87 per lead)
• Leads exclusive to him: roughly 8 of those 18
• Close rate on shared leads: 28%
• Close rate on exclusive leads: 44%
• Jobs booked from Angi per month: 5-7

Five to seven jobs from $1,400 and hours of chasing people who'd already talked to three other contractors. Kyle also ran Google Ads at $600/month, which added another 3-4 jobs. So his total was $2,000/month for about 9-10 booked jobs.

His average job was $7,500. That math works on paper. But it's fragile in practice — every dollar depended on platforms he didn't control. If Angi changed the lead pool or Google shifted his ads, his schedule emptied the next week.

He had no say in any of it. He was renting his business.

Why He Targeted Property Managers

Kyle had one type of client he loved more than any other: property managers.

They managed apartment complexes, small commercial buildings, HOA communities. When a roof needed work, they had the authority to approve $30,000 to $70,000 jobs without waiting on committee approval. They needed reliable contractors they could call year after year — someone already in their contacts when the problem came up, not a random name from a platform.

His theory was simple: if he could build direct relationships with 15-20 property managers in middle Tennessee, he'd have more pipeline than Angi had ever given him — and it would belong to him.

He'd tried cold outreach once before. He paid a virtual assistant $800 for one month to send emails on his behalf. She sent about 80 messages to a generic list.

He got two replies. Neither converted.

The idea wasn't wrong. The execution was bad. Generic emails to unqualified contacts with zero follow-up don't work. They also quietly damage your sender reputation.

How He Set Up the AI Outreach

A friend in Nashville — the owner of an HVAC company — told Kyle about an AI outreach tool that researched prospects individually, wrote personalized pitches, and followed up automatically. Not a blast of templates. An actual system that treated each contact like they mattered.

Kyle signed up in January 2025. Setup took about 45 minutes.

He told the tool:

• His service area: Murfreesboro plus a 40-mile radius
• Who he wanted to reach: property managers with 10+ units, commercial building owners, facility managers at restaurants and retail chains
• His angle: 15 years in the business, specialized in flat and low-slope commercial roofing, fast start times
• What to skip: residential jobs under $3,000, emergency repair-only requests

The system started with a warmup period — 10 to 15 emails per day to build sender reputation. Then it scaled to 20. Then 40.

Month One: Slower Than He Expected

January felt like nothing was happening. Kyle got four real replies all month. Not spam filters, not angry unsubscribes — actual responses.

Two had potential jobs in them. Neither closed.

He almost stopped. "Angi at least sends leads faster," he told himself.

But he did the math. Four qualified replies for $89 in tool cost. Versus 18 Angi leads for $1,400 — half of which went to competitors simultaneously, and most of which never answered his calls.

The speed comparison wasn't fair. Angi was selling him a feeling of activity, not actual results.

He kept going.

The Shift in Months Two and Three

February was different. The tool was sending 40-50 emails per day by mid-month. Kyle had given it feedback on which response types felt like real opportunities versus time-wasters. The messages got sharper as the system learned his preferences.

February by the numbers:

• Emails sent: approximately 900
• Replies received: 34 (about 3.8% reply rate)
• Qualified conversations: 18
• Estimates requested: 8
• Jobs booked from this channel: 3

Three jobs from $89 in marketing spend. All from people who had never heard of Kyle before January.

March changed the entire picture. A property management company he'd been in contact with for six weeks — three outreach messages plus one reply chain — called him about a full commercial roof replacement on a 12-unit complex. The job came in at $51,000.

That single project paid for years of subscription fees.

By the end of March, Kyle had 15 qualified leads active in his pipeline for the week. Not shared Angi prospects who'd already called his competitors. Real decision-makers who'd replied to him specifically, had read his specific pitch, and had asked for a follow-up.

What Made the Emails Actually Work

Kyle paid close attention to which messages got responses. Three things stood out.

First, specificity beat length every time. Emails that referenced something specific about the reader's situation — "managing multi-unit properties, you probably deal with flat roof maintenance every few years" — got replies at nearly twice the rate of generic versions. One accurate sentence is worth more than three paragraphs of generic pitch.

Second, the first email's only job is to start a conversation, not close a deal. Emails that tried to pitch too hard in message one killed the conversation before it started. The ones that worked asked a single, low-stakes question or made a relevant observation.

Third, follow-ups weren't optional — they were where most of the deals actually came from. Property managers are busy.

The first email might land at a bad time. The second lands when they're thinking about something related. The third one often catches them right when they need someone.

The Before-and-After Numbers

Before (Q4 2024):

• Marketing spend: $2,000/month
• Leads per week: 2-3
• Jobs per month: 9-10
• Close rate: ~32%
• Best monthly revenue ever: $72,000

After (April 2025, 90 days in):

• Marketing spend: $489/month ($89 AI tool + $400 Google Ads — Angi canceled in February)
• Leads per week: 13-15
• Jobs per month: 14-17
• Close rate: 48%
• Revenue that month: $124,000

The close rate jump makes sense when you think about what these leads actually are. The people who replied to Kyle's emails had already decided he was worth talking to. They weren't comparison-shopping three Angi bids. They'd read his message, thought "this guy sounds like what I need," and responded.

That's a different type of conversation from the start.

Kyle's Own Words

"The biggest surprise was how many property managers had been sitting on a roofing problem for a year or more. They just never found someone they trusted enough. My emails gave them an easy way to raise their hand — and I was already in their inbox looking like someone who knew commercial roofing. When they needed someone, they called me first."

He added something that stuck: the volume of leads matters less than most contractors think. He doesn't need 100 leads. He needs 15 good conversations with people who can approve a $30,000 job. That's the whole game.

One Thing He'd Do Differently

"Start in October, not January."

Roofing picks up fast in Tennessee during March and April. Property managers who need spring work start thinking about it in January. If Kyle had been sending emails since fall — warming up relationships through November and December — those contacts would've been ready to pull the trigger two months earlier.

"I left two months of spring pipeline on the table," he said. "The system works. But it's not instant. You need 60 to 90 days of runway before you hit full speed."

That's the thing most contractors miss. Outreach programs aren't a tap you turn on and get leads the same week. They're a system you build over time. The roofers who are fully booked in March started building their pipeline in November.

Kyle's advice: don't wait until business slows down to build the system. Build it now, before you need it.

What This Means If You're Still on Angi

Kyle still runs Google Ads at a reduced budget. Emergency calls — someone's roof is actively leaking — convert well from search because the urgency is immediate and the buyer is ready.

But for commercial work, the kind that defines his year? He's done with Angi. He sends emails directly to the decision-makers he wants to work with, and those relationships are his.

Three of the property managers he connected with in early 2025 have already sent him a second job without him doing any additional outreach. They put him in their "trusted contractor" folder and called him when the next project came up.

That's what owning your pipeline looks like.

---

If you're a roofer spending $1,000+ per month on shared leads, LeadClaw does the research, writes the emails, and follows up automatically while you're running jobs. Start free — no credit card required.